**4. Discussion**

In this pilot study, we found an improvement in overall symptoms of IBS and an improvement in the immune-inflammatory state and integrity of intestinal barrier. These aspects were especially significant in patients treated with *B. longum* ES1 for 12 weeks versus 8 weeks.

Accumulating evidence suggests that commensal bacteria may play a role in IBS and that specific probiotic therapy is able to improve GI symptoms in such patients [37]. Indeed, a recent systematic review showed that patients with IBS had lower microbial α-diversity in both stool and intestinal mucosal samples, with an overall microbial profile characterized by increased levels of Firmicutes and decreased levels of Bacteroidetes compared to healthy subjects [38]. On the other hand, there is a lack of consistency among the gu<sup>t</sup> microbiota fingerprints reported in different studies, probably due to different study designs, different methods used for bacterial profiling (including samples storage, DNA extraction and sequencing) and different statistical approaches [39]. In 48% of patients with IBS-D, it has been shown that probiotic therapy provided adequate relief of overall IBS symptoms, amelioration of stool consistency and a trend towards improved quality of life [40]. More recently, Giannetti et al. reported that the administration of a probiotic mixture of *Bifidobacterium infantis* M-63, *breve* M-16V, and *longum* BB536 in patients with IBS decreased abdominal pain frequency and improved QoL in a significantly higher proportion of patients, when compared with placebo [41]. Moreover, Pinto-Sanchez et al. showed that probiotic therapy (*B. longum* NCC3001) improved psychiatric comorbidities and increased quality of life in patients with IBS [42]. Consistently, we observed an improvement in IBS symptoms especially after 12 weeks of therapy with *B. longum* ES1. However, compared to previous studies, we did not observe an amelioration in the quality of life of our population. Likely, the probiotic mixture or specific strain beneficial for each patient remains to be determined.

Human and animal studies support the concept that a low-grade inflammation may perturb GI reflexes and activate the visceral sensory system contributing to the altered GI physiology and hypersensitivity underlying IBS [43,44]. Moreover, a recent study on murine model demonstrated the important role of neuron-glial communication mediated by TNFα and glial activation in visceral inflammatory hypersensitivity [45]. Consistently, we found a significant positive correlation between abdominal pain intensity and TNFα levels. Several studies have reported an imbalance in pro-inflammatory cytokines, such as IL-6, IL-8 and TNF<sup>α</sup>, in patients with IBS compared to controls [46–49]. Unfortunately, in the present study, we were not able to substantiate this data due to the lack of a healthy control population. Lastly, we found a positive correlation between IL-8 and TNF<sup>α</sup>, in accordance with the chemotactic and activation function performed by IL-8 against neutrophils, which in turn produce TNF<sup>α</sup>. Treatment with *B. longum* ES1 led to a significant reduction in cytokine levels from baseline to the end of therapy, contributing to the clinical improvement observed in our population. Taken together, these results confirm both the low-grade inflammation described in patients with IBS [36] and the anti-inflammatory effect of *B. longum* ES1 and, overall, support probiotic therapy as a valid treatment option in patients with IBS.

Finally, in the intricate puzzle of IBS pathophysiology, it has been shown that altered intestinal permeability plays an essential role, particularly in its diarrhoea-predominant variant. Indeed, several case-control studies showed that intestinal permeability measured by multi-sugar absorption test increased in patients with IBS-D compared to healthy controls [50,51]. Moreover, Linsalata et al. identified two different IBS-D subtypes that showed different inflammatory status according to intestinal permeability function, further supporting the concept that an impaired GI barrier may allow easier passage of luminal antigens that, in turn, may elicit an inflammatory response influencing the course of the disease [52]. In support of this issue, we found a positive correlation between serum zonulin, a marker of intestinal permeability, and IL-6, secreted by macrophages as a result of the Pathogen Associated Molecular Patterns (PAMPs) binding to the Toll-like receptors. Moreover, we observed a positive correlation between age and zonulin levels. A likely explanation for these findings could lie in a prolonged period of exposure to PAMPs. More recently, it has been reported that treatment with *B. longum* BB536 and *Lactobacillus rhamnosus* HN001 with vitamin B6, compared to

placebo, significantly improved GI symptoms and restored intestinal permeability in patients with IBS, as deduced by the improved percentage of sucralose recovery [53]. To date, very few studies have investigated serum zonulin concentration in patients with IBS. Singh et al. showed that serum zonulin levels in patients with IBS were higher compared with healthy controls and comparable to those with active coeliac disease [9]. Taking into account the results on serum zonulin concentration from previous studies conducted at our Center on IBD patients, we observed that serum zonulin concentration in patients with IBS-D was comparable to that found in patients with IBD (43.3 (95% CI 37.3–46.4) vs. 45.3 (95% CI 43.5–47.8) ng/mL, respectively) [30] and higher than those found in healthy controls (8.6 (95% CI 7.2–10.5) ng/mL) [8]. In the present study, a significant reduction in serum zonulin concentration was observed in patients who received probiotic therapy for 12 weeks and not in those treated for 8 weeks. These findings sugges<sup>t</sup> the presence of an altered intestinal permeability in IBS-D patients, and we observed an improvement of the barrier integrity in patients treated for 12 weeks.

The main limitations of the current study are the limited sample size and the lack of a placebo-treated control group, to provide definitive conclusions on the real e fficacy of *B. longum* ES1 treatment in our population. We observed a significant improvement of inflammatory status and stool consistency, irrespective of the treatment duration. None of the clinical outcomes allow any inference on the e ffect on IBS natural history in view of duration of therapy, nature of IBS symptoms (waxing and waning), and marginal di fferences. Since this was a pilot study, its aim was to evaluate whether our hypothesis could be su fficiently supported to justify a detailed investigation. Such data are promising and set the basis for controlled clinical trials with broader case studies.
